A simulation-based framework to optimize occupant-centric controls given stochastic occupant behaviour

Occupant-centric control (OCC) strategies represent a novel approach for indoor climate in which occupancy patterns and occupant preferences are embedded within sequences. They aim to improve comfort energy efficiency by learning predicting behaviour (OB), then optimizing building operations accordingly. Previous studies estimate that OCC can increase savings up 60% while improving comfort. However, their performance is subject several factors, including uncertainty due OB, configurational settings, as well design parameters. To this end, testing OCCs adjusting settings prior implementation critical ensure optimal performance. Furthermore, identifying alternatives optimize such given different an important step faces many logistical constraints during field implementations. This paper presents framework simulation environment, entails coupling synthetic OB with learn preferences. The genetic algorithm multi-objective optimization used identify the parameters minimize consumption maximize under various scenarios. demonstrate proposed framework, three were implemented program, EnergyPlus, executed through Python packages configurations Results revealed significant improvement of when they customized identified It was found these points could reduce 33% 28%, relative baseline scenario non-optimized implementation. aims actual buildings avoid discomfort issues may arise its initial phases.